[Eeglablist] [erplab] About stimulus presentation delay

Steven J Luck sjluck at ucdavis.edu
Wed Sep 21 15:00:16 PDT 2016


Hi everybody.  We are about to release a new version of ERPLAB that contains a routine for shifting event codes to deal with these delays.  Of course, you’ll need to measure the delays, but then it will be easy to shift the codes.

Steve


On Sep 21, 2016, at 10:04 AM, RICHARDS, JOHN <RICHARDS at mailbox.sc.edu<mailto:RICHARDS at mailbox.sc.edu>> wrote:

By the way, I am sure most of you must be familiar with the lag issue that EGI had with its Netstation amp.  An older version of the firmware in the amp was incorrectly calculating the lo-pass filter and all the points were lagged by a significant amount.  This did not show up in the EPrime-stimulus delay, but in the EGI-EPrime synch times.  The delay was significant re ERP component timing, so in this case any data that were collected with this machine should be adjusted.  Several people I know had moved from the older version to the newer version, and so had to adjust the delay to some data and not others.

I have cc’ed the support at EGI, and they can comment on this if they wish.  Also you can find info about this, I googled “EGI latency problem” and found a discussion of this, https://sccn.ucsd.edu/pipermail/eeglablist/2015/010427.html .  This has been extensively discussed elsewhere also.

John


***********************************************
John E. Richards
Carolina Distinguished Professor
Department of Psychology
University of South Carolina
Columbia, SC  29208
Dept Phone: 803 777 2079
Fax: 803 777 9558
Email: richards-john at sc.edu<applewebdata://7FD23341-F720-46A5-97FD-B4149780F9D4/richards-john@sc.edu>
HTTP: jerlab.psych.sc.edu<http://jerlab.psych.sc.edu>
*************************************************
<image001.png>

From: <erplab-request at ucdavis.edu<mailto:erplab-request at ucdavis.edu>> on behalf of John Richards <RICHARDS at mailbox.sc.edu<mailto:RICHARDS at mailbox.sc.edu>>
Date: Wednesday, September 21, 2016 at 11:10 AM
To: Colin Noe <cmn2rice at gmail.com<mailto:cmn2rice at gmail.com>>, chao wang <hnchaowang at gmail.com<mailto:hnchaowang at gmail.com>>
Cc: EEGLAB List <eeglablist at sccn.ucsd.edu<mailto:eeglablist at sccn.ucsd.edu>>, "erplab at ucdavis.edu<mailto:erplab at ucdavis.edu>" <erplab at ucdavis.edu<mailto:erplab at ucdavis.edu>>
Subject: Re: [erplab] About stimulus presentation delay

Nearly everyone has a delay between the EPrime time and the time the stimulus reaches the screen.  It should be measured.  The latency depends largely on the method that your computer uses to send a figure in memory to the display (< 1 frame) and the latency of the display.  If you are using a “projector” then this may be a long and variable time.  A “scope” is nearly instantaneous.  We use a computer monitor with a fast display rate, and our lags are < ½ video frame.  For audio, we use the recommended EPRime audio code, and our delays are about 3 ms with almost no variability.

If you have a constant or fairly constant lag, AND, if you are very concerned about using latency measures as a DV, then you should definitely adjust.  Otherwise, adjusting depends on the resolution of your ERP components.  We study the “Nc”, which is a rather slow-wave component in infants lasting from 350 to 700 ms.  So being off by < 1 frame (or even several frames) will not affect the mean or max ERP measures.  We also study the “N290” (an infant analog of the N170).  In this case the latency of the component is a DV in many studies, so adjusting is important.  The peak latency of the N290 is fairly variable across participants, so we could do the ERP analysis of peak ERP at peak latency (and mean around peak, or mean from nn to nn ms) and not be too badly off with no delay adjustment.  But since we want latency meausres also, we adjust.

If you have an extremely variable lag, say greater than the resolution of the ERP component you want to study, then you are out of luck.  The avg latency lag will be the latency lag for NO stimulus presentation, and the ERP will be noisy.  If the variablty of the stimulus lag is too great re your EEG/ERP analysis resolution, then you should fix your display.

John


***********************************************
John E. Richards
Carolina Distinguished Professor
Department of Psychology
University of South Carolina
Columbia, SC  29208
Dept Phone: 803 777 2079
Fax: 803 777 9558
Email: richards-john at sc.edu<applewebdata://7FD23341-F720-46A5-97FD-B4149780F9D4/richards-john@sc.edu>
HTTP: jerlab.psych.sc.edu<http://jerlab.psych.sc.edu>
*************************************************
<image002.png>

From: <erplab-request at ucdavis.edu<mailto:erplab-request at ucdavis.edu>> on behalf of Colin Noe <cmn2rice at gmail.com<mailto:cmn2rice at gmail.com>>
Date: Wednesday, September 21, 2016 at 10:49 AM
To: chao wang <hnchaowang at gmail.com<mailto:hnchaowang at gmail.com>>
Cc: EEGLAB List <eeglablist at sccn.ucsd.edu<mailto:eeglablist at sccn.ucsd.edu>>, "erplab at ucdavis.edu<mailto:erplab at ucdavis.edu>" <erplab at ucdavis.edu<mailto:erplab at ucdavis.edu>>
Subject: Re: [erplab] About stimulus presentation delay

I recognized the issue ahead of time, and programmed a delay into my marker code so that it lined up.  There is a way to move marks in the EEG though.

As for acceptability, that is a question for the journal editors and experts in this forum.

On Tue, Sep 20, 2016 at 9:12 PM chao wang <hnchaowang at gmail.com<mailto:hnchaowang at gmail.com>> wrote:
Thanks for your quick reply! So, how do you segment data with respect to stimulus onset?

My delay is a little complex, about 60% ~ 75% are within the average +/- 5 ms. I want to use the average value to segment EEG data or shift ERP component latency. But I am not sure if this is acceptable in ERP analysis?

Thanks!
Chao

On Wed, Sep 21, 2016 at 9:58 AM, Colin Noe <cmn2rice at gmail.com<mailto:cmn2rice at gmail.com>> wrote:
I have a jitter of +- 3 ms with mine (along with a 50 ms delay), which I assume is better than most (who likely have not checked timing).

On Tue, Sep 20, 2016 at 8:40 PM chao wang <hnchaowang at gmail.com<mailto:hnchaowang at gmail.com>> wrote:
Hi, everyone,

We're recording EEG data and we use E-Prime to present stimuli. During recording, stimuli are presented to subjects via a projector. The projector induces a delay between stimulus presentation on E-Prime computer and in front of subjects. We measured this delay and found that the delay was not constant across trials but instead jitters within certain range around the average. I would like to know is there a typical range of the jitter that can be accepted in ERP experiments? And in this situation, what should I do to segment EEG data with respect to stimulus onset? I mean, use the average delay to determine the real instant of stimulus presentation, or any other ways exist? Forgive me if my words is not clear :)

Thank you in advance!

Chao


--------------------------------------------------------------------
Steven J. Luck, Ph.D.
Director, Center for Mind & Brain
Distinguished Professor, Department of Psychology
University of California, Davis
Room 109
267 Cousteau Place
Davis, CA 95618
(530) 297-4424
E-Mail: sjluck at ucdavis.edu<mailto:sjluck at ucdavis.edu>
Web: http://mindbrain.ucdavis.edu/people/sjluck
Calendar: http://www.google.com/calendar/embed?src=stevenjluck%40gmail.com&ctz=America/Los_Angeles
--------------------------------------------------------------------

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20160921/6edb36f7/attachment.html>


More information about the eeglablist mailing list